Background: Radiotherapy (RT) is one of the main treatments for prostate cancer (PC). The effectiveness of RT depends on the initial radioresistance of tumor cells provided by their certain molecular features - the transcriptional activity of genes, miRNAs and long non-coding RNAs (lncRNAs). Together, these indicators form a competing endogenous RNAs network (ceRNAs). The aim of this work was to study the ceRNA features in PC patients with different responses to RT. Methods: The study included 400 patients with PC, 90 of them developed biochemical recurrence after RT (Novalis TX, TFD = 75 Gy). Tumor tissue samples were obtained by biopsy. The RNA fraction was isolated and purified using the RNeasy Plus Universal Kits. Bioinformatic search for miRNAs was performed using a modified TarPmiR algorithm. The miRNA-lncRNA interactions were assessed by bioinformatics. Expression of AKT, ATM, BRIP, BRCA1/2, CDK1, CDKN1B, CCND1/3, EXO1, H2AX, KU70, PTEN, RAD50, RAP80, RIF1, RNF168, TOPB1, TP53, XRCC4, BAX, CASP-8, -3, -9, MDM2, BCL2, RBBP8, EP300 genes, miRNA and lncRNA were determined by RT-qPCR. Differences were assessed using the Mann-Whitney test (Bonferroni correction was used). Results: The expression of CDK1, CDKN1B, RBBP8, XRCC4, BRCA2, RAD50 genes in tumor tissues of patients with biochemical recurrence (group 1) was higher (p < 0.005) and the expression of TP53 and BCL2 was lower than the values in tumor tissues in patients without recurrence (group 2). Bioinformatic search for miRNAs targeting the CDK1, CDKN1B, RBBP8, XRCC4, BRCA2, RAD50, TP53 and BCL2 revealed 1725 miRNAs, 78 of which were validated (miRDB) and had a minimum free energy of miRNA-mRNA interaction. Patients in group 1 showed differential expression of miRNAs targeting genes BCL2 (increased miR-139-5p, miR-4446-5p, miR-211-5p); CDK1 (decreased miR-330-3p, miR-5580-3p, miR-429), CDKN1B (decreased miR-6841-3p); RAD50 (decreased miR-433-3p, miR-3691-5p); RBBP8 (decreased miR-130b-3p); TP53 (increased miR-6884-5p, miR-149-3p, miR-6860, miR-4728-5p, let-7b-5p); XRCC4 (decreased miR-4489) and BRCA2 (decreased miR-7151-3p). The interaction of these miRNAs with 160 lncRNAs was mathematically predicted. Patients in group 1 demonstrated increased expression of lncRNAs LINC00641, SCAMP1, MALAT1, MATN1-AS1 (miR-429), LINC00657 and GAS5 (miR-433-3p), CCAT1 (miR-130b-3p) and reduced expression of SNHG3, DHRS4-AS1 (miR-139-5p) and MAL2 (let-7b-5p). Conclusions: Our study of the ceRNA network features and RT effectiveness for PC established the mechanisms of radioresistance development and its predictors.